a trademark of Seagate Removable Storage Systems, LLC. Other product names
are trademarks or registered trademarks of their owners.
Seagate reserves the right to change, without notice, product offerings or
specifications. No part of this publication may be reproduced in any form without
written permission from Seagate Technology, LLC.
Publication Number T001, Rev. C; February, 2001
Important Information About This Manual
All information contained in or disclosed by this document is proprietary to Seagate
Technology. By accepting this material, the recipient agrees that this material and
the information contained therein are held in confidence and in trust and will not be
used, reproduced in whole or in part, nor its contents revealed to others, except to
meet the purpose for which it was delivered. It is understood that no right is
conveyed to reproduce or translate any item herein disclosed without express written
permission from Seagate Technology.
Seagate Technology provides this manual "as is," without warranty of any kind,
either expressed or implied, including, but not limited to, the implied warranties of
merchantability and fitness for a particular purpose. Seagate Technology reserves
the right to change, without notification, the specifications contained in this manual.
Seagate Technology assumes no responsibility for the accuracy, completeness,
sufficiency, or usefulness of this manual, nor for any problem that might arise from
the use of the information in this manual.
Page
iv
LTO SCSI Interface Manual
Contents
Introduction1
About This Manual...................................................................................................... 1
SCSI Interface Overview2
ANSI SCSI Bus Standards ................................................................................... 2
Cabling and Connectors....................................................................................... 2
Completion Status ............................................................................................ 126
IntroductionPage
Introduction
The Seagate LTO drives are designed for computer environments requiring high
performance, ultra-high capacity data storage. LTO drives are available in 5.25-inch
internal configurations or as an external subsystem.
Seagate LTO drives contain an embedded Small Computer Systems Interface
(SCSI) controller. These drives provide synchronous or asynchronous SCSI and a
high speed burst synchronous data transfer rate of 80 Mbytes per second. The
internal drive is tailored for easy installation in today's computers and the fullfeatured embedded SCSI controller facilitates easy integration into a variety of
systems.
About This Manual
1
The information in this manual applies specifically to the Viper 200 drive. For drivespecific installation and operational information, including SCSI connection
information, refer to the Installation Guide or Product Description Manual for the
specific model of your drive.
Page
2
SCSI Interface Overview
The Seagate LTO drives are designed to operate with the Small Computer System
Interface (SCSI) bus. This chapter discusses SCSI bus operation as it pertains to
drive functions.
SCSI is a standard interface established to support peripheral equipment such as
printers, tape drives, magnetic disks, optical disks for microcomputers and other
computer systems. The SCSI bus can support up to sixteen devices consisting of
any multiple of host adapters and peripheral devices.
The Seagate LTO drives comply with SCSI-2 and SCSI-3 specifications. In a few
cases, vendor unique features are available. These features are compatible with the
SCSI standards.
The interface is a sixteen-port daisy-chained bus using twenty-seven signal lines:
eighteen data-bit signal lines and nine control lines. The eighteen data-bit lines are
made up of two pairs of eight data signal lines and one parity bit line. The remaining
nine lines provide control and status signals to coordinate data transfer operations
between the host controller and the selected drive.
LTO SCSI Interface Manual
The drives have an internal SCSI controller integrated into the drive electronics.
Each device ID on the SCSI bus may drive up to 16 logical units (LUN). Seagate
LTO drives support only LUN 0.
ANSI SCSI Bus Standards
In addition to the information presented in this manual, we recommend that you
review the applicable SCSI-3 standards before writing host software drivers. Also,
see the conformance statements, which are given in the Product Description Manual
for the each model of LTO drive.
Cabling and Connectors
The cabling requirements and pin assignments for the SCSI connector for the
internal drive models are given in the respective installation guide and Product
Description Manual for each model of LTO drive.
SCSI Interface OverviewPage
SCSI Conformance
General features
The Viper 200 supports the following:
• Disconnect/reconnect, arbitration (required in SCSI-2 and SCSI-3)
• LVD and single-ended drivers
• Both single and multi-initiator systems
• Fixed and variable block transfer lengths
• Hard reset
• Synchronous data transfers
• Parity implemented (switch option)
• Space blocks, filemarks and EOD
• Log Sense and Log Select for managing soft error reporting
Interface standards conformance
3
The Small Computer System Interface is described in standards including several
versions and a number of individual documents. The original Small Computer
System Interface Standard, X3.131-1986, is referred to herein as SCSI-1. SCSI-1
was revised resulting in the Small Computer System Interface – 2 (X3.131-1994),
referred to herein as SCSI-2. The set of SCSI-3 standards are collectively referred to
as SCSI-3. The applicable SCSI-3 standards are as follows:
SCSI-3 Architecture Model (SAM) X3.270: 1996
SCSI Architecture Model - 2 (SAM-2) in development
SCSI-3 Parallel Interface Amendment (SPI Amnd) X3.253/AM1: 1998
SCSI Parallel Interface - 2 (SPI-2) X3.302: 1999
SCSI Parallel Interface - 3 (SPI-3) in development
SCSI-3 Primary Commands (SPC) Status: X3.301: 1997
SCSI Primary Commands - 2 (SPC-2) in development
SCSI-3 Medium Changer Commands (SMC) NCITS.314: 1998
SCSI-3 Stream Commands (SSC) in development
The term SCSI is used wherever it is not necessary to distinguish between the
versions of SCSI.
LTO Conformance
Seagate LTO drives conform to the requirements of the
Command Set Advisory Document
Generation 1 SCSI Common
(AU1-SCSI) for Ultrium format devices.
Page
4
Strategy
Reset Condition
Unit Attention Condition
LTO SCSI Interface Manual
When a reset is issued to the tape drive, the SCSI bus clears all uncompleted
commands, releases all SCSI device reservations, sets the tape drive to default
modes, and returns to the Bus Free phase.
The Unit Attention condition in the tape drive typically results from the following
conditions:
• A Reset was previously issued to the drive.
• The drive has just been powered on.
• A cartridge has been inserted.
• A log exception condition occurred.
• The mode pages have been changed.
• The cartridge has been loaded or unloaded to the Hold position, and the Medium
Auxiliary Memory (MAM) is accessible.
The Unit Attention Condition persists for each Initiator until that Initiator issues a
command other than Inquiry for which the tape drive returns with a Check Condition
Status. If the next command from that Initiator following the Check Condition Status
is Request Sense, then the unit attention sense key is returned.
If the Inquiry Command is received from an Initiator with a pending Unit Attention
Condition before the tape drive reports Check Condition Status, the tape drive
performs the Inquiry Command and does not clear the Unit Attention Condition.
If the Request Sense Command is received from an Initiator with a pending Unit
Attention Condition before the tape drive reports Check Condition Status, the tape
drive reports unit attention sense key and clears the Unit Attention Condition for that
Initiator.
Contingent Allegiance Condition
The Contingent Allegiance condition shall exist following the return of Check
Condition and may exist following an unexpected disconnect. The contingent
allegiance condition shall be preserved for the I T x nexus until it is cleared. The
contingent allegiance condition shall be cleared upon the generation of a hard reset
condition or by an Abort message, a Bus Device Reset message, or any subsequent
command for the I T x nexus. While the contingent allegiance condition exists the
drive shall preserve the sense data for the initiator.
While the contingent allegiance condition exists, the drive shall respond to any other
requests for access to the logical unit from another initiator with a Busy status.
Execution of queued commands shall be suspended until the Contingent Allegiance
condition is cleared.
SCSI Interface OverviewPage
Buffered Mode
Buffered Mode allows the most efficient operation of a tape drive. The drive defaults
to Buffered mode. In this mode, the drive signals Command Complete when all
requested data for a Write command has been transferred from the host to the tape
drive buffer. This mode provides data to maintain operation while the host readies a
new Write Command.
If an error occurs in writing data to the tape after the tape drive signals Command
Complete, an error status is sent on the next Command issued.
Immediate Function
For Initiators that do not support the disconnect feature, the Immediate bit provides a
means of releasing the bus while the drive is busy completing a function such as
repositioning the tape. If a command is sent by the Initiator after a previous
Immediate Command was accepted, the drive continues the Immediate Function it is
currently performing and may disconnect, execute, or reject the new command
depending on the new command.
An immediate bit of zero means that the status is returned to the Initiator when the
operation is completed. (For example, the status is returned when the tape has been
repositioned.) An Immediate bit of one means that the status is returned to the
Initiator as soon as the function is started.
5
Residual Length Function
When performing a Write command, the drive returns a Good Status and Command
Completion Message when the last byte requested by the command is placed in the
Data Buffer, rather than when it is written onto tape. If an error occurs while data is
being written onto tape, the drive calculates the Residual Length and places this
value in the information bytes of the Sense Data Block. Also, the Residual Length
functions for other commands, such as Read.
Residual Length is calculated by: RL = TL - AL
Where:
AL (Actual Length) =Blocks transferred from the host to the tape drive across
TL (Transfer Length) =The Transfer Length from bytes 2-4 of the Write com-
RL (Residual Length) =The amount of blocks or bytes not written to tape.
the SCSI bus.
mand (Request Transfer Length).
Page
6
Disconnect/Reconnect Function
When the drive is performing a task not requiring communication with the Initiator or
when the tape drive determines that a relatively long time has passed with no bus
activity, it disconnects from the SCSI bus. Examples are:
• When rewinding the tape.
• When writing to the tape and the buffer is full.
• When reading from the tape and the buffer is empty.
• When spacing, locating, or generally performing any tape motion when data
cannot be transferred on the SCSI bus.
During the time the Target is disconnected for one of these functions, the bus is free
for use by other devices. Both disconnect and reconnect are initiated by the Target.
If the tape drive is selected while disconnected, it only allows the following actions:
• If the command is from a different initiator or is from the same initiator but to a
different LUN, the tape drive accepts the command and immediately disconnects
if the command is a media-access command. If the new command is a Request
Sense, Inquiry or Test Unit Ready, then the new command is executed
immediately.
• Immediately following the selection, the Initiator may send the Identify, No Op,
Abort, or Bus Device Reset messages to the drive.
• If the command is from the same initiator to the same LUN, the current
command terminates with a Check Condition and an Abort Sense Key.
LTO SCSI Interface Manual
Early Warning Function
Early Warning on the Viper drive is a logical warning given when 64 megabytes of
storage space remain on the tape. The position is calculated by the drive. When this
physical position is reached on a tape, the following occurs.
1.Data transfers from the host are terminated at the next disconnect burst size
boundary for SEW=1.
2.All data remaining in the drive buffer is written to the tape if SEW=1.
3.The command completes with a Check Condition and a 40h Sense data mean-
ing EOM and no Sense Key.
4.Subsequent WRITE commands write data and complete with check condition
with EOM Status and No Sense Key until the physical tape end is encountered.
Error Reporting
Soft errors are generally tape-quality related and occur more frequently during write
operations than during read operations. Soft errors indicate repeated attempts by the
drive to read or write data on the tape. Some soft errors are normal, but an increase
in the usual count can indicate deteriorating tape quality. If the soft error count
SCSI Interface OverviewPage
remains higher than normal, use a cleaning cartridge to clean the read/write heads. If
this procedure does not clear the problem, change to a new tape cartridge.
If a hard error (unrecoverable error) occurs during operation, the drive terminates
operation immediately and returns a Check Condition. The Initiator should cease any
further read or write functions and issue a Request Sense Command to determine
the type of error.
When the drive detects a write error, it attempts to rewrite the data up to 128 times.
After the 128th attempt, the error is considered unrecoverable and the operation
terminates. When a hard error is encountered, replace the tape with a new cartridge
and repeat the function or clean the heads with a cleaning cartridge.
Variable and Fixed Mode Recording
The Viper drive can write either fixed or variable block sizes. The recording mode is
determined by the Fixed bit in the SCSI Write and Read commands.
If the Fixed bit is set, the Mode Select command sets the size of the next block or
multiple blocks to be written with the next Write command. When a Write command
is issued with the Fixed bit set, the current block size is implemented. The transfer
length specifies the number of blocks to be written with this size. If a Write command
is issued with the Fixed bit set and the current block size set to 0, the Viper drive
returns a Check Condition with Illegal Request Sense Key. When writing with the
Fixed bit set, each Write command specifies the number of contiguous blocks to be
written of a fixed size, resulting in fixed-mode blocks. If the Fixed bit is reset, then
only one block can be written on the tape per SCSI Write command CDB. In that
case, the Write command CDB transfer length specifies the size of the block to be
written in bytes. With the Fixed bit reset, the current block size specified with the last
Mode Select command is ignored.
7
Setting the block size to 0 in the Mode Select page descriptor is not required.
Therefore, with the Fixed bit reset, each SCSI Write command may specify a
different byte count, resulting in variable-mode blocks.
The host may switch between fixed and variable mode recording. By issuing the
Mode Select command to specify different block sizes, blocks can be written to the
tape with different block sizes in the fixed mode. Also, the host may change the block
size after BOM, allowing on-the-fly block-size changes.
The Read command Fixed bit also specifies fixed or variable mode. When reading in
variable mode, the host must know the size of the block to be read from the tape in
advance in order to avoid causing the Viper drive to return a Check Condition with
Incorrect Length indicated in the Sense data (ILI). Also, the data transfer may be
truncated (cut off) when the recorded block does not match the transfer length in
variable mode or the current block size in fixed mode.
The Read command includes a SILI bit to Suppress ILI Check conditions.
When the SILI bit is set, the host usually specifies the maximum block size before
reading so that the data blocks are not truncated, and no Check Conditions are
generated.
Page
8
Library Features
Auto Unload Mode
LTO SCSI Interface Manual
The SCSI Read Block Limits command returns the minimum and maximum block
sizes that the Viper drive can support. The Block Limits data is not modified to reflect
the current mode of writing—fixed or variable. The Block Limits returned data is not
modified to reflect the current block size for the next fixed-mode Write. The Mode
Sense command is used for that purpose.
The Viper drive has an Auto Unload Mode that may be accessed via a Mode
command to Mode Page 21h. Auto Unload mode allows the drive to be configured
for Library systems or as a standalone drive. Auto Unload mode allows the drive to
control how the tape is unloaded in automated environments. Auto Unload modes
provide three options for tape unloading:
• Do Nothing – Stay at BOT or Stay SEATED.
• Unthread Only
• Unthread and unload
The list below provides examples of automation situations in which the ejection
process may be controlled using the Auto Unload mode. These examples include not
only normal power-on situations, but situations involving cleaning tapes, incompatible
tapes and firmware tapes.
• Powering on or power cycling while the tape is inside the drive
• Cleaning tape has finished the cleaning cycle
• Unsupported data tape has been loaded
• Unsupported cleaning tape has been loaded
• Microcode download from a firmware tape
• Microcode download via the SCSI Write Buffer command while a tape is loaded
SCSI ID Modification
Tape libraries require that the SCSI ID of an installed tape drive can be changed
without opening the library enclosure and changing address jumpers. To change the
SCSI ID of the Viper drive, issue a Mode Select command to Mode Page 22h to
change the Next Selection ID field to the desired SCSI ID. After you reset or cycle
the power to the drive, it will begin responding to the new SCSI ID. The new ID will
persist through additional resets and power cycles.
If the address jumpers are changed and the drive’s power is cycled, the drive will
respond to the new address on the jumpers. The value in the Next Selection ID field
will be that on the jumpers, even though the field may have been changed
immediately before the power cycle.
LUN 1 Command Forwarding
A tape library or autoloader containing a Viper can receive its SCSI commands via
the Viper. This function is called “Command Forwarding” and is enabled via the
“CmdFwd” field in the Interface Control mode page (22h). When enabled, commands
received via the SCSI bus and addressed to LUN 1 are immediately transmitted to
SCSI Interface OverviewPage
the library via the RS-422 interface. The Viper acts as a bridge, passing commands,
data, and status between the initiator and the library, as specified in the Library
Interface Encapsulated SCSI Protocol Manual.
In SCSI Medium Changer terminology, the library or autoloader appears as an
“independent medium changer,” because the tape drive and medium changer are at
different LUNs. Medium changer commands addressed to LUN 0 are rejected with
Check Condition and sense data of Illegal Request / Invalid Command Operation
Code (05/20/00).
Because of the relatively slow communication on the RS-422 interface, the Viper will
disconnect from the SCSI bus after receiving a command, after receiving data-out,
and after sending status. Data-in and status are sent to the initiator without
disconnecting between them.
If the link between the Viper and the library fails, a command will receive a status of
Check Condition and sense data of Aborted Command / Logical Unit Communication
Failure (0B/08/00).
If a SCSI bus reset is received, a target reset task management request is
transmitted to the library.
9
When command forwarding is disabled in mode page 22h, commands addressed to
LUN 1 will receive a status of Check Condition and sense data of Illegal Request /
Logical Unit Not Supported (05/25/00).
After Command Forwarding is enabled or disabled, the drive will respond to the next
command with a Check Condition and sense data of Unit Attention / Reported LUNs
Data Has Changed (06/3F/0E).
Page
10
SCSI Messages and Status Codes
SCSI Messages
The SCSI message codes, descriptions, and directions are given in the following
table. Each of these SCSI messages are supported by the Seagate LTO drive.
In = Drive to host
Out = Host to drive
* Only two extended messages are Supported: Synchronous Data Transfer Request and Wide Data
Transfer Request.
LTO SCSI Interface Manual
Inbound messages
The following table lists drive to host SCSI messages and their definitions.
MessageHex CodeDefinition
Command
Complete
Disconnect04hThis message is sent by the drive to indicate t hat it is about to dis-
Extended
Message
Identify80h+An Identify in message is sent to the init i ator during reconnect to in-
Ignore Wide
Residue
00hThis message is sent by the drive at the end of the s t atus phase to
indicate that a command i s complete. Once the message is sent, the
drive releases the bus and goes t o Bus Free
connect from the bus and go to B us Free. During a Data phase, it is
always preceded by a Save Data Pointers message. If a Message
Reject message is rec ei ved in response to this message, then the
disconnect is prevent ed.
01hSee page 11 for a description of ex t ended m essage support
dicate which Logical Unit i s reconnecting. The Disconnec t Privilege
and LUNTAR flags will both be clear
23hThis message is sent by the drive to the initiator to indicate that a
byte on a wide bus is not valid. This will be s upport ed whenever a
wide transfer is act i ve.
SCSI Messages and Status CodesPage
MessageHex CodeDefinition
Message
Reject
Restore
Pointers
Save Data
Pointers
07hThis message is sent t o the initiator when the mess age received by
the drive is unsupported or inappropriat e.
03hThis message causes the initiator to reset its data transfer pointers
to the values they held when t he l ast Save Data Pointers m essage
was sent. It will be sent when a parit y error is detected on the bus or
when an Initiator Detected Error message is received in order to retry the data phase.
02hThis message instruc t s the initiator to save its current data trans fer
pointers for use with a subs equent Restore pointers mess age. Thi s
message will always be sent before a Disc onnect message during
data phases.
Outbound messages
The following table lists host to drive SCSI messages and their definitions.
MessageHex CodeDefinition
Abort06hAn abort condition is generated. See Message exception handling on
Bus Device
Reset
Extended
Message
Identify80h+The Identify Out mess age i s sent by the initiat or to identify the Logical
Initiator
Detected
Error
Message
Parity Error
Message
Reject
No
Operation
0ChA reset c ondi tion is generated. . See Message exception handling on
01hSee Ex tended messages on the followi ng page for a description of
05hThe initi ator has detected an error in the data being sent in a Message
09hThe initi ator has detected a parity error in a m essage. The drive will
07hThis message is sent when the initiator does not support a mes s age
08hThis message has no effect and is ignored.
page 14.
page 14.
extended message support
Unit to be accessed and to set Disconnect Privilege. The LUNTAR flag
must be zero.
The Identify Out mess age m ust be sent as the firs t thing after selection. If it is sent at any other time, the drive will respond with a mes sage reject message and go to B us Free.
Command, Data or Status phase. The drive will retry the data burst or
message.
If the message is rec ei ved immediately after an Identify message or
after the Command Complete mes sage has been sent, the drive wil l
go Bus Free.
retry the message.
If the message is rec ei ved during a Command, Data or Status phase,
immediately after an I dent i fy message or after the Com m and Com plete message has been sent, the driv e will go B us Free.
sent by the drive or consi ders the message inappropriate. If the message being rejected is Disconnect, Synchronous Data Transfer Request or Wide Data Transfer Request, the operation continues wit hout
those features. For all ot her m essages except Restore Pointers, the
message is treated as an A bort message.
If the message is rec ei ved during a Command, Data or Status phase,
immediately after an I dent i fy message or after the command complete
message has been sent, the drive will go B us Free.
11
Page
12
Extended messages
The following table lists extended SCSI messages and their definitions.
Hex CodeMessageDefinition
Synchronous
Data Transfer
Request
Wide Data
Transfer
Request
Parity errors
LTO SCSI Interface Manual
01hThe drive will never initiate a Synchronous dat a t ransfer negotia-
tion, but will expect the initiator to do so.
If the message is rec ei ved after selection and bef ore c ommand
phase, it will then go to Message In phas e and res pond with a valid
response to complete the negot i ation.
If the message is rec ei ved at any other time, a Message Reject is
sent in response.
03hThe drive will never initiate a Wide Data Transfer negotiati on, but
will expect the initiator to do so.
If the message is rec ei ved after selection and bef ore Com m and
phase, it will then go to Message In phas e and res pond with a valid
response to complete the negot i ation.
If the message is rec ei ved at any other time, a Message Reject is
sent in response.
The following table describes the operation for each of the possible cases of parity
error.
State or PhaseDescription
Bus Free StateThe drive does not detect nor reac t to parity errors on the SCSI bus while
Arbitration PhaseThe drive does not detect nor react to parity errors on the SCSI bus while
Selection PhaseThe drive does not detect nor react to parity errors on t he SCSI bus while
Selection, Message Out
Phase (Identify Mes sage)
Reselection, Mess age I n
Phase (Identify Mes sage)
Command PhaseIf the drive detects a pari ty error while the host is transferring a CDB, the
the drive is in a bus free state.
arbitration is being performed.
the drive is being select ed.
If the drive detects a parity error while the host is sending an Identify
message, the drive retries forever.
If the drive is att em pting to reconnect to the hos t and the host asserts
ATN because it detect ed an error, the drive:
♦
Switches the host t o the Message Out Phase.
♦
Waits for the host to send a 09 (Parity Error Message)
♦
The drive then performs the retry option by:
♦
Switching the host to the Message In Phase.
♦
Resending the Identify Message
drive:
♦
Switches the host t o M essage In phase
♦
Sends Restore Data Pointers (03h) m essage
♦
Switches the host t o Com m and phase to retry the command.
♦
If a parity error occurs on the retry the host:
♦
Terminates the transfer.
♦
Switches the host t o the Status Phase and sends a Check Condition.
♦
Switches the host t o the Message In Phase and sends a Command
Complete.
♦
Sets the Sense Key = B and ASC/ASCQ = 47-00 (SCSI-2 only).
SCSI Messages and Status CodesPage
State or PhaseDescription
Data In PhaseIf the host detects a parity error while data is being transferred from the
Data Out PhaseIf the driv e det ects a parity error while the host is transferring data, the
Status PhaseIf the host is in t he S tatus phase and detects an error i n the status byte
Message In PhaseIf the host is in t he M essage In Phase and detects an error on a message
Message Out Phase
drive and asserts ATN, t he dri ve:
♦
Terminates the transfer of dat a.
♦
Switches the host t o the Message Out Phase.
♦
Waits for the host to send a 05 (Host Detected Error).
♦
Switches the host t o the Status Phase and sends a Check Condition.
♦
Switches the host t o the Message In Phase and sends a Command
Complete.
♦
Sets the Sense Key = B and ASC/ASCQ = 48-00 (SCSI-2 only).
drive:
♦
Terminates the transfer of dat a.
♦
Switches the host t o the Status Phase and sends a Check Condition.
♦
Switches the host t o the Message In Phase and sends Command
Complete.
♦
Sets the Sense Key = B and ASC/ASCQ = 47-00 (SCSI-2 only).
and asserts ATN, the dri v e:
♦
Switches the host t o the Message Out Phase and waits for the host to
send 05 (Initiator Detect ed E rror).
♦
Switches the host t o the Status Phase and sends a Check Condition.
♦
Target sends restore pointers and resends Status.
bytes an asserts A T N, the drive:
♦
Switches the host t o the Message Out Phase.
♦
Waits for the host to send 09 (Parity Error Message).
♦
Switches the host t o the Message In Phase and resends the message.
♦
If the host is in t he M essage Out Phase and sends a mes sage bytes
and the drive detects a parit y error, the drive retries forever.
13
An Initiator that accommodates disconnect/reconnect can indicate this capability to
the tape drive during the Selection phase by asserting both its own Initiator SCSI ID
bit as well as the tape drive’s SCSI ID bit (allows the tape drive to know with which
Initiator to reconnect). The Initiator must also assert ATN before exiting the Selection
phase (prior to releasing SEL) and send an Identify message out of C0h to the tape
drive. This sequence causes the drive to enter the Message-Out phase when the
Selection phase completes.
The first message sent by the host after the Selection phase is an Identify message.
Under normal conditions, the first message sent by the tape drive after a Reselection
phase is also Identify. Under certain exceptional conditions, the host may send the
Abort message or the Bus Device Reset message instead of Identify as the first
message.
Page
14
LTO SCSI Interface Manual
Message exception handling
Message exceptions are handled as described in the following tables. The numbers
in the tables identify the actions that the drive takes when an exception under the
identified conditions occurs. The actions are listed following the tables.
StatusStatus has been reported
No Status Status has not been reported
Identify
(80h)
If retry, 4,
else 5
If Retry, 4,
else 5
CC
(00h)
2If retry, 4,
2If Retry, 4,
Ignore
Wide Residue
(23h)
else 5
else 5
else 4
else 4
SCSI Messages and Status CodesPage
Actions:
1. The Message Out Reject disables the feature and continues. In case of Message
In Disconnect, the drive disables further “Disconnects” on the current command.
In case of Message In SDTR, both the initiator and the drive go to asynchronous
mode. In case of Message In WDTR, the drive goes to 8-bit data transfer mode.
2. The drive goes Bus Free by releasing BSY without sense data set up.
3. The Message Out Abort or Bus Device Reset will result in the drive going Bus
Free and therefore abnormally terminating the command.
4. The drive re-sends the message.
5. The drive posts Check Condition status indicating an Aborted command.
6. The drive rejects the message by Message In Reject and continues its operation.
7. The drive rejects the message by Message In Reject.
8. The drive sends Message In Restore Pointers and retries if retry is not
exhausted.
15
9. The drive ignores the message and continues.
10. The wide data transfer agreement is negotiated prior to negotiating the
synchronous data transfer agreement.
11. The drive goes Bus Free by releasing BSY with sense data set up.
12. The drive continues the normal process.
13. If the drive is continuing an operation after sending a Message in Reject, the
drive will post Check Condition status indicating an Aborted command (Action 5),
if not, The drive will go Bus Free by releasing BSY without sense data set up
(Action 2).
Page
16
SCSI Commands
This chapter describes the SCSI commands for the LTO tape drive.
Command Set
The following table shows the SCSI commands for sequential access devices
implemented by the drive.
GroupCodeCommandPage
000hTest Uni t Ready116
001hRewind110
003hRequest S ense98
005hRead Block Limits79
006hPark Unpark79
008hRead73
00AhWrite120
010hWrite Fi l em arks125
011hSpac e114
012hInquiry22
013hVerif y118
015hMode Sel ect42
255hMode Sel ect42
016hReserv e Uni t108
017hRelease Uni t91
019hErase20
01AhMode Sense64
25AhMode Sense64
01BhLoad Unload33
01ChReceive Diagnostic Results89
01DhSend Diagnostic111
01EhPrevent/Allow Medium Removal70
00BhSet Capacity112
12BhLocate36
134hRead Pos i t i on84
13BhWrite Buffer122
13ChRead Buffer81
244hReport Dens i t y Support93
24ChLog Select38
24DhLog Sense42
256hReserv e Uni t108
257hRelease Uni t91
3A0hReport LUNS96
LTO SCSI Interface Manual
SCSI CommandsPage
Conventions
The commands in this chapter are listed in alphabetical order. Each command is
described, its Command Descriptor Block (CDB) illustrated, and the Completion
Status is given. Bits and fields defined in the ANSI SCSI documents that are not
used by the drive are not described in this document. Bits and fields that are
supported by the drive are described.
Command Descriptor Blocks
A host makes request of the tape drive by sending a Command Descriptor Block
(CDB). Some commands also require a parameter list. If the CDB or the parameter
list contains an invalid parameter, the drive terminates the command, returning a
Sense Key of Illegal Request, without altering the medium.
Command Descriptor Block Formats
SCSI six-byte Command Descriptor Blocks are arranged in the format shown in the
following table.
Field names that are centered represent fields that are at least one byte long. Field
names that are left-aligned with a bit number are one or more bits long. The length is
indicated by the beginning of the following field name, and by the Bits column in the
Field Descriptions table, shown below.
Page
18
Field Descriptions
FieldByteBitsDescription
Operation
Code
Ignored
Logical
Unit
Command
Dependent12-4
ReservedAll reserved bits mus t be 0.
ControlLastThe Control byte is made up of t he Vendor Unique, Flag and Link
Fields that are one byte or longer have no entry in the Bits column.
Operation Code Format
Byte
0
LTO SCSI Interface Manual
0The Operation Code is made up of the Group Code and the Com-
mand Code. See Operation Code Format, below.
15-7The LUN field is res i dual from the SCSI-1 days and is i gnored.
0-4See the specific command.
2-8
bits. See Control Format, below.
Bits
76543210
Group CodeCommand Code
Field Descriptions
FieldByteBitsDescription
Group
Code
Command
Code *
NOTE: The Group Code and Command Code together make the Operation Code.
Control Byte Format
Byte
Last
Field Descriptions
FieldByteBitsDescription
Vendor
Unique
ReservedLast2-5All reserved bits mus t be 0.
Flag and
Link bits
05-7The SCSI comm and group.
00-4The SCSI comm and code.
Bits
76543210
Vendor UniqueReservedFlagLink
Last6-7Not used, always 0.
Last0-1These bits are not supported by t he Viper drive.
SCSI CommandsPage
Command Status
When the tape drive competes a command it responds with a status byte. The format
of the status byte is shown below.
Status Byte Format
Bits
Byte
0
Field Descriptions
FieldByteBitsDescription
Reserved06-7Always 0
Status
Code *
Status Codes
76543210
ReservedStatus Code
00-5The command status, as shown below.
19
Status codes for the Viper drives are shown in the following table.
DescriptionValueWhen returned
Good Status00hThe command completed without problems
Check Condition02hA problem occurred during command execution. The sense data
Busy08hThe drive is unable to accept the command at t hi s time. This is
Reservation Conflict18hThis is returned if the drive is reserved for an initiator other than
should be examined to determine t he nat ure of the problem.
only returned during the power on sequence or if there are commands from too many init i at ors outstanding.
the one sending the command
Page
20
ERASE
Command Descriptor Block
The Erase command marks the tape, from the current position to the end of tape, as
erased.
Byte
0
1
2
3
4
5
Field Descriptions
FieldBytesBitsDescription
Operation
Code
Ignored
Logical
Unit
Immed110 = Status is not returned until the tape is erased.
Long100 = S hort Erase is performed and EOD is recorded at the current
ReservedAll reserved bits mus t be 0.
ControlSee Control Byte Format on page 18.
LTO SCSI Interface Manual
Bits
76543210
Operation Code (19h)
Ignored LUN ReservedImmedLong
Reserved
Reserved
Reserved
Control
0The Operation Code for Erase is 19h.
15-7The LUN field is res i dual from the SCSI-1 days and is i gnored.
1 = Status is returned as soon as the operation is initi ated.
tape position.
1 = Long data-security Erase i s performed, and EOD is recorded
from the current tape posit i on t o the end of the tape.
If the Immediate (Immed) flag is set to 1, then the drive validates the command and
waits for any previous command from any host to complete including any immediate
commands currently being processed and for any buffered data to be flushed to
tape. It will then report a deferred error for any preceding command or buffered data
if appropriate. If there is no deferred error, the drive reports good status and initiates
the command. If the Immediate (Immed) flag is set to 0, status is not returned until
after the command has completed.
If the Long bit is set, EOD is written at the current position. Data Set Separators are
then written from EOD to the end of the medium to overwrite any data currently on
the tape. If the Long bit is clear, then an EOD is written at the current position
marking it as end of data.
SCSI CommandsPage
Completion Status
CodeMessageDescription
00h
02h
Good Status
Check
Condition
•
•
Note: If Immed is 1, then Good Status only indi cates that the comm and i s
valid.
Use the Request Sense comm and to retrieve status information. Possibl e
Sense Keys are:
CodeMessageDescription
02hNot ReadyNo cartridge is in the drive.
04hHardware ErrorParity error on the SCSI bus or drive hard-
05hIllegal RequestBoth the Immed bit and t he Li nk bits are 1, or
06hUnit AttentionThe cartridge was changed or the drive was
07hWrite ProtectThe cartridge is write protected.
The drive remains in any previous l y set modes.
The drive is ready to perform any appropri at e command.
ware failure detected.
the Command Descriptor Bloc k is invalid.
reset prior to this command.
21
Loading...
+ 106 hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.